1. Introduction
1.1 Background
Leishmaniasis is a vector-borne parasitic disease affecting up to one million people annually, mostly neglected populations living in rural areas with poor access to health care (Burza, Croft, and Boelaert 2018). It exists primarily in two forms: visceral leishmaniasis, which is the most severe manifestation of the disease and can lead to death, and cutaneous leishmaniasis, which is the most common manifestation of the disease and is associated with social stigma as well as depression, anxiety, and other mental health conditions (Burza, Croft, and Boelaert 2018). The main endemic areas for visceral leishmaniasis are Eastern Africa, South Asia, and Latin America (mostly in Brazil), while cutaneous leishmaniasis is principally found in Latin America, the Middle East, and North Africa. Treatments for leishmaniasis presents the limitations of being parenteral drugs with risk of severe side effects. The only oral drug available, miltefosine, has gastro-intestinal side effects and, more importantly, is potentially teratogenic (Dorlo et al. 2012). This necessitates WOCBP to use mandatory contraception during and after treatment for approximately five months. Furthermore, research is currently testing new oral short treatments for leishmaniasis, but as they need to be assessed for reprotoxicity risks, the use of contraception is also required during these clinical trials. This represents a substantive access barrier for WOCBP to both treatment and participation in clinical trials, which respectively limit access to the best treatment available, and limit evidence on effectiveness and safety of potentially teratogenic drugs such as miltefosine among WOCBP. Therefore, there is a need to inquire about acceptability rate of contraceptive use when medically indicated alongside treatment for leishmaniasis, and by extrapolation, for NTDs at large.
There is no available empirical estimate of the prevalence of leishmaniasis in West Pokot. However, the number of visceral leishmaniasis cases treated in Kacheliba Sub-County Hospital, which was the only centre providing treatment for visceral leishmaniasis at that time, was 645 in 2022. Regarding contraceptive use, modern methods rate was estimated at 23.2% in 2022, with implants (11.7%) and injections (8.3%) being the most used methods in this county – unmet needs were estimated at 30.3% (Kenya National Bureau of Statistics, Nairobi, Kenya, Ministry of Health Nairobi, Kenya, and The DHS Program, ICF, Rockville, Maryland, USA 2023a). This latter rate was estimated at 13.3% in 2014 (Kenya National Bureau of Statistics et al. 2015). Some of the most salient reasons hindering the use of contraceptives in this region have been reported to be strongly related to religious and cultural beliefs around family planning, and low access to health facilities (Oscar Kakai 2022). The total fertility rate was estimated at 5.3 children per women in West Pokot in 2023, while it was 3.3 in Kenya in 2022 (Kenya National Bureau of Statistics, Nairobi, Kenya, Ministry of Health Nairobi, Kenya, and The DHS Program, ICF, Rockville, Maryland, USA 2023b; County government of West Pokot 2022). The National Council for Population Development (NCPD) suggests that this high fertility rate fuels poverty and food insecurity (Oscar Kakai 2022).
Requiring to use contraceptives has also been reported as a barrier to WOCBP’s participation in clinical trials (Couderc-Pétry et al. 2020; Milliez 2009). In a Phase 3 clinical trial in Eastern Africa, the non-inferiority of a paromomycin and miltefosine regimen was assessed against a sodium stibogluconate and paromomycin combination for treating visceral leishmaniasis. Authors noted that the low participation of women relative to men is likely to be due to the requirement of using contraceptives, especially in West Pokot (Musa et al. 2023).
This report consists in an intermediary overview of the quantitative results of a partcipatory mixed methods project about medically indicated contraceptive use conducted in West Pokot, Kenya. These results will be used to inform the qualitative strand of this project in two ways. First, it will provide information about the sampling of the participants to the in-depths interviews (IDIs) - for instance, a profile of interest with regards to the research objectives of this project is women who reports intentions to use contraceptive for family planning matters but not alongside treatment. Selecting women displaying such an intention pattern to participate to IDIs could provide the research team the possibility to inquire about the precise reason explaining this specific pattern. Secondly, the results of this quantitative analysis will provide avenues to refine the content of the qualitative tools. After the qualitative data will have been collected and analyzed, quantitative and qualitative data will be integrated to provide an holistic understanding of rate of acceptability of medically indicated contraceptive use as well as the reasons for use and non-use, on which recommendation will be framed in collaboration with members of the West Pokot community.
1.2 Objectives
General objective
- To assess the acceptability of WOCBP to use modern contraceptives when medically indicated during standard treatment for leishmaniasis as well as during clinical trials in West Pokot, Kenya.
Specific objectives
What is the rate of intention to use contraceptive when medically indicated and in the context of clinical trials for the treatment of leishmaniasis among WOCBP visiting the Kacheliba and Sigor Sub-County Hospitals?
Is there a difference in intention to use modern contraceptive methods when medically indicated for leishmaniasis treatment versus future intention to use contraceptive for family planning among WOCBP visiting the Kacheliba and Sigor Sub-County Hospitals?
What factors, if any, are associated with utilization of modern conceptive methods when recommended alongside medical treatment, as well as in the context of medical treatments?
2. Methods
2.1 Study design and period
This study consists in the quantitative phase of a mixed methods research project on the topic of medically indicated contraceptive use. The project relies on a explanatory sequential design, and its quantitative strands follows a hospital-based cross-sectional design implemented through a structured survey. It was administered between the 21 and 25 October 2024, and collects information on the topic of intention to use contraceptives when indicated alongside medical treatment for leishmaniasis. Before being triangulated with the qualitative results, the findings presented in this report will be used in order to refine tools related to the qualitative phase of the study (i.e. in depth interview guide and focus-group discussion guide), as well as inform the sampling of in depth interview participants.
2.2 Study area
The questionnaire was administered in the only two health centers delivering treatment for leishmaniasis situated in West Pokot, Kenya: Kacheliba and Sigor District Hospitals.
2.3 Inclusion and exclusion criteria
Inclusion criteria
- WOCBP aged 15-49 years old living in endemic area covered by the Kacheliba Hospital or Sigor Sub-County Hospital, irrespective of their contraceptive use and/or pregnancy status.
Exclusion criteria
- Women who, in the view of their treating physician (or the investigator if medically qualified), are not medically fit to participate.
2.4 Sample size calculation and sampling procedure
The required sample size was estimated using a single population proportion formula, which result was adjusted using a finite population correction to account for the population size.
As no quantitative data about contraceptive use acceptability in the context of medical treatment have been published in this region, an estimated proportion of 50% for contraceptive use acceptability when recommended for medical reasons was used to maximize variance. Confidence interval and the margin of error were set at 95% and 5% respectively. These calculations yield a sample size of 384 participants, to which a finite population correction has been applied based on the Kacheliba and Sigor catchment populations, which respectively are 18,090 and 14,172 (facilities data, unpublished). For both figures, 21.4% was taken to estimate the sampling frame, which corresponds to the proportion of women of childbearing potential in the catchments, yielding a result of 6904. This correction resulted in an adjusted sample size of 364, to which a 10% non-response rate was added, resulting in final sample size of 401 participants. The formula mentioned here in the above are the following:
Single population proportion formula
\(x = \frac{Z_{\frac{\alpha}{2}}^2 \cdot p \cdot (1-p)}{\text{MOE}^2}\)
Finite population correction formula
\(n = \frac{N \cdot x}{x + N - 1}\)
where Zα/2 is the critical value of the Normal distribution at α/2 (i.e. confidence interval), p is the estimated sample proportion, MOE is the margin of error and N is the sampling frame.
2.5 Data collection instrument
A structured and pre-tested interview-based questionnaire was used to collect data, which includes closed-ended and open-ended questions. The questionnaire was administered on tablets by research assistants using the Kobo collect app. This instrument is available in Appendix I.
2.6 Data processing and analysis
Data were entered, cleaned, coded and analysed using Rstudio, version 2024.04.02 Build 764. Descriptive and analytic statistical procedures were applied.
Analytic statistical procedures began with a McNemar test of independence for paired proportions to examine whether intentions to use contraceptives differed depending on their context of use—specifically, in the context of medical treatment (e.g., alongside Kala-azar treatment) versus family planning. Subsequently, three models were constructed to run bivariate and multivariate logistic regressions, with outcomes including: intentions to use contraceptives when medically indicated alongside Kala-azar treatment, future intentions to use contraceptives for family planning, and ever usage of contraceptives. After verifying the absence of multicollinearity, these models included variables selected based on the theoretical framework guiding this study to ensure alignment with its conceptual foundations. To complement this theoretically driven approach, a forward stepwise selection using the Akaike Information Criterion (AIC) was applied to each of the three models. The results of this data-driven modeling were not directly presented in this report but were incorporated into the discussion to provide additional insights and to contrast with the theoretically driven findings
2.7 Quality control
Co-investigators introduced the data collection instrument to research assistants (RAs), and trained them to use them. RAs were given time to familiarize themselves with the instrument by performing role plays among temselves, and subsequently during the piloting of the tool. During data collection, daily meeting took place to provide a space for RAs to communicate to their peers and the co-investigators about their daily experience in administering the questionnaire, and voice any questions or issues encountered during the day.
2.8 Operational definitions
The instrument use to collect data analysed in the subsequent section is availble in appendix I. The description of variables which have constructed is found just below.
CU Intention alongside treatment
Intentions to use contraceptive alongside medical treatment for Kala-azar. A short prompt was provided to inform participants about a potential situation where she could asked if she would be willing to use contraceptives in this context (see Appendix I for the prompt). This variable is evaluated using 4 modalities, in the form of a likert scale (i.e. strongly likely, likely, unlikely, strongly unlikely). To conduct a logistical regression using this variable as outcome, this variable has been binarized as follows:
“Some intentions”: participants who answered “Strongly likely” or “Likely” to the CU intention alongside treatment variable.
“No intention”: participants who answered “Unlikely” or “Strongly unlikely” to the CU intention alongside treatment variable.
CU intention during clinical trials
Similar question to the previous one, but in the context of a clinical trial and not during standard treatment. As for standard treatment, a prompt was provided to respondents to give some context (see Appendix I for the prompt). The same scale was used for the answer’s modalities than for the above question.
Future intentions to use FP
This variable captures future intentions of participants to use contraceptives in the context of family planning. This was evaluated using a 4 point likert scale, similarly to the two variables described above. To conduct a logistical regression using this variable as outcome, this variable has been binarized as follows:
“Some intentions”: participants who answered “Strongly likely” or “Likely” to the “Future intentions to use FP” variable.
“No intention”: participants who answered “Unlikely” or “Strongly unlikely” to the “Future intentions to use FP” variable.
Ever used
This variable was created to capture whether respondent ever used contraceptives throughout their life. It is build from two questions, one asking about participants current use of contraceptives, and another one asking respondents about their past use. This latter question was only asked to respondents who reported not using any contraceptive method currently. Participants who reported either using a contraceptive method currently, or used some in the past, where assigned a “Yes” to the ever used variable, while other were assigned a “No”.
Economic status
Economic status has been built from two variables, one asking about respondent’s monthly income, and in case respondents did not now their monthly income, they were asked about their monthly spending. The final economic status variable includes three modalities: low (monthly income < KSH 23,670 or monthly spending < KSH 4,999), medium (monthly income above KSH 23,670 but less than KSH 200,000, or monthly spending between KSH 5,000 and 49,999) and high (monthly income equal or above KSH 200,000, or monthly spending equal or above 50,000).
Knowledge of Kala-azar
Knoweldge related to Kala-azar has been built from different questions related to Kala-azar means of transmission and symptoms. It includes four modalities:
“No knowledge”: participants who reported not knowing Kala-azar as a disease.
“Transmission knowledge”: participants who know at least one mean of transmission of Kala-azar.
“Symptom knowledge”: participants who know at least one symptom related to Kala-azar.
“Complete knowledge”: participants who know both at least one symptoms and means of transmission associated with Kala-azar.
Experience of Kala-azar
This variable classifies participants based on how much they have been in contact, or heard about, people sick with Kala-azar. It includes three modalities:
“Seen/heard of sickness”: participants who report having seen or heard about someone with Kala-azar infection.
Seen/heard of death”: participants who report having seen or heard about someone dying from Kala-azar infection.
“No contract”: participants who neither have seen/heard about someone infected or who died of Kala-azar.
Feared treatment-related consequences
Number of consequences associated with Kala-azar treatment feared by participants (e.g. pain related to injections, prolonged hospital stay, anxiousness/depression). This variable includes two categories:
“No/low concerns”: participants who reported fearing not more than one treatment related consequence.
“Moderate/high concerns”: participants who expressed fearing more than 1 treatment related consequences.
Knowledge of contraceptive methods
Number of contraceptive methods known by participants, which includes 3 modalities:
“Low knowledge”: participants who reported knowing between 0 and 2 contraceptive methods.
“Moderate knowledge”: partcipants who reported knowing between 3 and 4 contracptive methods.
“High knowledge”: participants who reported knowing more than 4 contraceptive methods.
Fear of side effects
This variable follows the framing tested and validated by Zan & Rossier (2024). It rests on two questions: one related to the number of side effects known by respondent and the number of these that are feared. The number of side effects feared is then divided by the number of side effect known, which yields a scale varying between 0 and 1 (0 expressing no concerns, and 1 expressing the highest level of concern). Participants who reported knowing no side effects where considered as having no fear (i.e. a value of 0 for this variable). This continuous indicators has then been classified in three modalities for better understanding:
“Low fear”: participants whose responses yield a result ranging from 0 to 0.33.
“Moderate fear”: participants whose responses yield a result ranging from 0.34 to 0.66.
“High fear”: participants whose responses yield a results ranging from 0.67 to 1.
Contraceptive use approval
This variable follows the framing tested and validated by Zan & Rossier (2024). It is build on three questions related to the approval of the use of contraceptive within a couple in different context: to prevent birth, to space birth and to limit birth. These questions are evaluated on a 5 points likert scale which are assigned a value between 1 and 5 (with “Strongly agree” being assigned a value of 5 and “Strongly disagree” a value of 1). The value of the three questions have then been summed up for each participant, and thus yields a value varying between 3 and 15. This continuous indicator has then been framed in the following three modalities:
“Low approval”: participants whose answers yield a value varying between 3 and 6.
“Medium approval”: participants whose answers yield a value varying between 7 and 10.
“High approval”: participants whose answers yield a value above
Contraceptive use agency
This variable follows the framing tested and validated by Zan & Rossier (2024). It has been framed following the same process than for the contraceptive use approval variable described above (i.e. the values associated to the likert scale modalities are summed up to form a single value) - however, this variable rests on 5 questions related to contraceptive use agency, yielding a single value varying between 5 and 25. This continuous indicator has then been framed in the following three modalities:
“Low agency”: participants whose answers yield a value varying between 5 and 13.
“Medium agency”: participants whose answers yield a value varying between 14 and 19.
“High approval”: participants whose answers yield a value above
Decision maker
This variable has been framed from two variables asking about the ability of respondents to decide using contraceptives for themselves. The first one ask about whose taking such a decision (e.g. herself, their partner, herself and their partner jointly). Respondents who answered “Respondent and partner jointly” where then ask whose opinion between themselves and their partners matters the most in such decisions. The final decision maker variable includes two modalities:
“Respondent”: participants who expressed that they are taking decision related to their own use of contraceptives themselves, or, if respondents expressed that this decision is taken jointly, that their own opinion matters more or equally to the one of their partner concerning this topic.
“Husband or someone else”: participants who reported that their partner of someone else is taking decision related to their own use of contraceptives, or, if respondents expressed that this decision is taken jointly, that the opinion of their partner matters more than their own concerning this topic.
3. Results
Descriptive analysis
A total of 401 WOCBP were interviewed from Kacheliba and Sigor District Hospitals. Socio-demographic characteristics are found in Table 1.
Table 1
Socio-demographic variables
Concerning the age of participants, close to half of the sample is between 18 and 24 years old, and the teenager category is the least represented modality within the sample (21 participants, 5.2%). Most participants gave birth to 1 or 2 children (n = 166, 41.4%), which tends to be quite in line with the age distribution of the sample (i.e. 43.1% of participants have between 18 and 24 years old).
A majority of participants are living in rural areas (92.3%), sedentary (65.8%), and of Christian confession (96.8%). Furthermore, Almost half participants did not complete any level of formal education (48.1%) , and a quarter did complete primary education (25.7%). Most participants were considered as having a low economic status (81%) or medium economic status (18.2%).
Knowledge and perceptions of leishmaniasis
Knowledge of leishmaniasis was quite high among participants, with 33.4% of the sample knowing both at least a mean of transmission and one symptom, and 29.4% of participants who either never heard about leishmaniasis, or did not know any symptom nor means of transmission. Most participants had some experience with leishmaniasis, with 56.6% of the sample having heard or seen someone sick from leishmaniasis, and 21.7% heard or seen someone die from this disease. Furhtermore, participants showed relatively low concerns about consequences of leishmaniasis treatment, with above 70% of participants evaluated as having no or low concerns. Pain at the injection site (31.7%) and prolonged hospital stay (26.9%) were the most frequent concerns mentioned by participants.
Contraceptive use variables
Approximately a third of participants reported using contraceptives when interviewed (34.7%), while this rate jumps to 55.9% when reporting about ever usage of contraceptives. Figure 1 describes the kind of contraceptives methods reported to be used by participants who ever used contraceptives. Three quarter of participants reported being the decision makers related their use of contraceptives (75.1%), and 10.7% reported to have been pressured to give birth at some point in the past.
Figure 1
More than three quarter of participants reported being strongly likely (48.1%) or likely (30.7%) to use contraceptives for family planning in the future. In the context of clinical trials, 81% of participants reported being either strongly likely (50.1%) or likely (30.9%) to choose a treatment regiment that requires to use a contraceptive method during treatment over one that do not require contraceptive use, provided that the former treatment is less disruptive (in terms of number of injections, length of hospital stay and side effects) - this rate jumps to 85.1% (strongly likely: 58.4%, likely: 26.7%) in the context of standard treatment (i.e. outside clinical trials).
Binary analysis: difference in proportions
In order to establish whether a significant difference exists between participants reporting future intentions to use contraceptives for family planning and intentions to use some contraceptive methods in the context of standard treatment for leishmaniasis, a McNemar test was applied these two proportions. To do so, both of these variables were binarized - as shown in Table 2, both were originally recorded using a 4 point likert-scale. The results of this test are found in Table 3.
Table 3
The rows in Table 3 corresponds to intentions to use contraceptives for family planning purposes, while the columns corresponds to intentions to use contraceptives alongside medical treatment. What can be taken out of this table is that a strong and significant difference exists between the proportion compared (McNemar Chi-Squared = 8.67, p-value = 0.0032). The absolute difference, which indicates the net difference between the two proportions, further confirm the direction and significance of this difference. Future intentions to use contraceptive for family planning are 6.5% lower than intention to use contraceptives alongside Kala-azar treatment. The confidence interval [-10.9%, -2.2%] indicates that the true difference in proportions is unlikely to include 0, further supporting the significance of the result.
Multivariate analysis
Table 4-6 display three logistical regression models that differ structurally because of their respective outcomes: Table 4 includes intentions to use contraceptives alongside medical treatment as outcome, Table 5 includes future intentions to use contraceptives as outcome, while Table 6 includes ever use of contraceptive as outcome. These models have been set up in order to compare the difference in predictors for each outcome, thereby allowing to identify difference in factors influencing contraceptives use intentions (and behavior, for ever usage) in these different contexts (for medical reasons and for family planning purposes). At the bottom of each table, the Akaike Information Criterion (AIC), which balances model’s complexity and goodness-of-fit, and the McFadden R-squared, which indicates the part of the outcomes’ variance explained by the predictors, are reported to assess overall model performance.
Table 4
Concerning the psychosocial predictors included in the model described in Table 4, all categories related to approval and agency with regards to contraceptive use are negatively associated with intention to use contraceptives alongside medical treatment. When taking into account contraceptive use approval for instance, respondents who showed low approval were 90% less likely than respondent showing high approval to hold intentions to use contraceptives alongside medical treatment. Fear of side effects is also positively associated with the outcome (although in a less significant manner as shown by the higher p-value): respondents showing low and medium fear of side effects were respectively 2.48 and 2.82 more likely to have intentions to use contraceptives alongside medical treatment than women displaying high fear of side effects. This result might seem surprising given that the odds of holding intentions to use contraceptives alongside treatment can be expected to be higher among the low fear category than the medium fear category - explanatory pathways of this unexpected results are covered in the discussion section. Similarly, ever use of contraceptives is also positively associated with the outcome. However, it is important to note that the ability of participants to decide using contraceptives for herself (i.e. the “decision making” variable) is not associated with intention to use contraceptives alongside medical treatment, as shown by the odds ratio and p-value. Furthermore, knowledge of contraceptive shows no significant association with intention to use contraceptives alongside medical treatment.
For sociodemographic predictors, no age category is significantly associated with the outcome, while only one category of the number of birth variable (i.e. 5-6 children) is positively associated with the outcome. No other sociodemographic variable were associated with intention to use contraceptive alongside treatment in this model.
Regarding disease related predictors, : as the number of concerns related to treatment consequences, intentions to use contraceptives for medical reasons increase. This is not surprising, since the treatment regiment for Kala-azar which does not involve using contraceptives (i.e. sodium stibogluconate and paromomycin combination) presents more negative consequences than the regiment requiring to use contraceptives (i.e. paromomycin and miltefosine). This precision was stated when prompting the question related to intention to use contraceptives alongside medical treatment. Knowledge of the Kala-azar appears however not be associated with the model’s outcome.
Table 5
Table 5 summarize the output of a logistic regression model which includes future intentions to use contraceptives for family planning purpose as outcome. This means that in this case, intentions to use contraceptives are still the phenomenon of interest, but outside a purely medical context. It also excludes disease-related predictors that were include in the model output in Table 4, as they are not relevant to contraceptive outside a medical context. Three predictors included in the model show significant associations with the outcome: number of birth(s), contraceptive approval and contraceptive use decision making ability.
Only two modalities of the parity variable significantly predict the outcome, namely 1-2 children and 3-4 children, which indicates that after controlling for the effects of other predictors, women who have between 1-2 and 3-4 children were respectively 4.27 and 6.04 times more likely to report intentions to use contraceptives for family planning purposes than nulliparious women. Regarding approval of contraceptive use, women showing low and medium approval were 88% and 64% less likely to report future intentions to use contraceptives for family planning purposes than women showing high approval. This shows a negative association between contraceptive use approval and the outcome. Lastly, women who reported taking decisions related to their own use of contraceptives themselves were 3.37 times more likely to report having future intentions to use contraceptives for family planning purposes than women who stipulated that their husband (or someone else) usually takes decision on this matter. In relation to this, it is interesting to note that contraceptive use agency is not significantly associated with the outcome in this model. This contrast with the results displayed in Table 4 where the opposite was the case: while contraceptive use agency was positively associated with intentions to use contraceptives alongside medical treatment, the decision making variable was not significantly associated with this latter outcome. Potential avenues explaining this difference will be described in the discussion section below.
Table 6
Table 6 reports the output of a model which includes the same predictors as the ones included in the model related to Table 5, but with ever usage of contraceptive as outcome instead of future intentions to use contraceptives for family planning purposes. This means that in this particular case, intentions are not the phenomenon of interest, but an actual behavior. In this context, 6 predictors showed associations with the outcome: age, lifestyle, education level, knowledge of contraceptive methods, contraceptive use approval and contraceptive use decision making ability.
Regard sociodemographic predictors, women between 30 and 39 years old were 4.67 times more likely to report having ever used contraceptives than the reference category (i.e. women between 15 and 17 years old) after accounting for other predictors. Women reporting a sedentary lifestyle also displayed higher odds to have ever used contraceptives than nomadic women. Furthermore, education level was also positively associated with ever use of contraceptive, with the strength of the associations incrementally increasing as the level of formal education increases: women who completed primary education were 1.91 times more likely to reported having ever used contraceptives, and women who completed secondary education or higher were 2.85 times more likely to report having ever used contraceptives, than women who did not attend the formal education system.
Concerning contraceptive use related predictors, women displaying low knowledge about contraceptives use were 68% less likely to report having ever used contraceptives than women with high knowledge on this topic, after adjusting for other predictors. Medium knowledge level category was however not associated with ever usage of contraceptives. Similarly, women showing low approval of contraceptive use were 26% less likely to reported having ever used contraceptives compared to women displaying high approval - medium approval was not significantly associated with the outcome. Finally, as in the case of the results displayed in Table 5, while contraceptive use agency was not associated with ever usage, women who reported having the ability to decide on their own whether or not to use contraceptives for themselves were 2.69 times more likely to have ever used contraceptives than women whose husband (or someone else) takes this decision.
Discussion
Overall, the findings of the quantitative phase of this mixed methods project demonstrate on the one hand that acceptability of contraceptives in a medical context (i.e. alongside standard treatment and during clinical trials) is higher than both contraceptives use behaviors (i.e. current use and ever usage) and future intentions to use contraceptives for family planning. On the other hand, multivariate analyses (Table 4, 5 and 6) further demonstrate that among the sample surveyed in this study, different sets of predictors are associated with different contexts of contraceptive use. Indeed, while predictors that have shown to be often associated with contraceptive use (e.g. education level, decision making ability) are actually also associated with contraceptive use in our model (especially in terms of ever usage but also for future intentions to use contraceptives for family planning purposes), such predictors are not significantly associated with intentions to use contraceptives alongside medical treatment.
Some possible avenues to be discussed:
difference between the models
agency and decision maker varies (decision_maker not significant in medical context but significant for FP - agency strongly associated in medical context but not really in FP context)
“Classic” predictors of contraceptive use are not so much associated with the medical context, in comparison to FP context.
“issue” with the fear of SE variable: output of the forward stepwise selection
8% of the sample fears issues related to fertility when using FP
drinking unboiled milk have been often reported as a cause of leishmaniasis
Most of participants expressing having some intentions to use contraceptives in medical context would be willing to do so for a period of up to 6 months